Genome‐wide association studies in Crohn's disease: Past, present and future

Abstract

Over the course of the past decade, genome‐wide association studies (GWAS) have revolutionised our understanding of complex disease genetics. One of the diseases that has benefitted most from this technology has been Crohn's disease (CD), with the identification of autophagy, the IL‐17/IL‐23 axis and innate lymphoid cells as key players in CD pathogenesis. Our increasing understanding of the genetic architecture of CD has also highlighted how a failure to suppress aberrant immune responses may contribute to disease development – a realisation that is now being incorporated into the design of new treatments. However, despite these successes, a significant proportion of disease heritability remains unexplained. Similarly, most of the causal variants at associated loci have not yet been identified, and even fewer have been functionally characterised. Because of the inarguable rise in the incidence of CD in regions of the world that previously had low disease rates, GWAS studies will soon have to shift from a largely Caucasian focus to include populations from other ethnic backgrounds. Future studies should also move beyond conventional studies of disease susceptibility into phenotypically driven ‘within‐cases’ analyses in order to explore the role of genetics in other important aspects of disease biology. These studies are likely to include assessments of prognosis and/or response to treatments and may be critical if personalised medicine is ever to become a reality.

Genome-wide association study of gastrointestinal disorders reinforces the link between the digestive tract and the nervous system

“…we do not report results of the MHC region (25Mb -34Mb) 48 . 14 15 MAGMA (v1.06) 46 (Multi-marker Analysis of GenoMic Annotation) was used to test for gene-based association 16 based on the SNP association results of the six digestion phenotypes. Gene length boundaries were defined as 17 35 kilobase (kb) upstream and 10 kb downstream from start and stop site, respectively, to include regulatory 18 elements.…”

“…For comparison, we also conducted IVW-MR 98 , MR-Egger 99 , 13 weighted median-MR 100 and MR-PRESSO 101 analyses following the STROBE-MR guideline 102 . 14 Genetic risk score (GRS) prediction 15 We used MD GWAS summary statistics 49 (European ancestry, excluding UKB cohort) as discovery data to 16 predict GP+M risk (risk for GORD, PUD and likelihood for taking GORD, PUD drugs). The MD data SNPs were 17 matched with the GP+M SNPs (7,156,547 SNPs), then LD pruned and "clumped", discarding variants within 18 1,000kb of, and in r 2 ³ 0.1 with, another (more significant) marker using SNPs with MAF > 0.01 from 10,000 19 random sampled unrelated UKB European-ancestry individuals as the LD reference.…”

Benchmarking network propagation methods for disease gene identification

“…It has been suggested that using genetic association information is one of the best ways to identify such drug targets [2]. In recent years, a large number of highly powered GWAS studies have been published for numerous common traits (see for example [3] or [4]) and have yielded many candidate genes. Further potential targets can be identified by adding contextual data to the genetic associations, such as genes involved in similar biological processes [5,6].…”

A Novel Approach to Detecting Epistasis using Random Sampling Regularisation

“…A common approach from big data techniques is to use multiple testing adjustments, as discussed later on in Methodology. Successes in GWAS have previously outlined viable SNPs in complex diseases such as Crohn's Disease [29], Rheumatoid Arthritis [30][31][32][33] and Celiac Disease [34]. It has also previously been proposed that GWAS studies should be a first step in the genetic identification process [43].…”